Differences in mortality from coronary artery bypass graft surgery at five teaching hospitals

Objective. —To measure hospital- and surgeon-specific mortality rates for patients with coronary artery bypass graft (CABG) surgery and to examine possible reasons for any differences. Design. —Cohort study using hospital discharge abstracts and itemized bills. Setting. —Five major teaching hospitals in Philadelphia, Pa. Patients. —Consecutive sample of all 4613 patients over a 30-month period. Main Outcome Measure. —In-hospital mortality rates. Results. —We observed differences in hospital mortality rates for patients who underwent coronary artery catheterization and CABG surgery during the same admission (diagnosis related group 106) but not for patients who underwent only CABG surgery during the admission (diagnosis related group 107). There were threefold differences in surgeon-specific mortality rates. The hospital mortality rates for coronary artery catheterization and CABG surgery during the same admission changed during the study and coincided with moves of surgeons among study hospitals. Our measures of illness severity did identify patients who were more likely to die, but differences in severity of illness did not explain differences in hospital- or surgeon-specific mortality rates. Patient mortality rates were not associated with the volume of procedures performed by individual surgeons. We found inconclusive evidence for an association with surgeons' clinical skills, and, to a lesser extent, with the hospital's volume of procedures and the hospital's organization and staffing. A greater intensity of hospital services was not necessary for a lower mortality rate. Conclusions. —We conclude that studies of CABG mortality should examine mortality rates by diagnosis related group, collect data from more than 1 year, examine associations with surgeons' clinical skills, include information on hospital organization and staffing, and cautiously explore more efficient ways of providing care. (JAMA. 1991;266:810-815)